experiment-craft

Debug a failed research experiment using a 5-step diagnostic flow: collect failure cases, identify a working baseline, isolate the cause, hypothesize and verify, and propose fixes. This structured approach traces root causes and reduces wasted iteration cycles.

Maintain structured experiment logs by recording Purpose, Setting, Results, Analysis, and Next Steps for each cycle. Using a standardized template ensures consistent documentation and supports cross-cycle learning across research iterations.

Trace the root cause of an underperforming method by applying a repeatable diagnostic workflow: collect the specific failure cases, identify a known working version, isolate variables, hypothesize the root cause, and verify it empirically before proposing fixes.

Yes, you can use this diagnostic workflow for cross-cycle research handoffs. The structured experiment logs document Purpose, Setting, Results, Analysis, and Next Steps, supporting seamless transfer of artifacts to downstream tasks like paper writing or review.

You need the experiment log template to document your debugging process. It requires recording Purpose, Setting, Results, Analysis, and Next Steps, ensuring systematic iteration logs are captured for diagnosing failing experiments and generating insights.

Use a 5-step diagnostic flow for experiment debugging when you need to diagnose failing experiments, trace root causes systematically, and maintain structured iteration logs for cross-cycle learning. It provides a disciplined approach to speed insight generation and reduce wasted cycles.